Choose Your Plan
Ship better on-device models with more confidence
Free Trial
For getting a taste
$0
- Access to real devices in our lab (Mac, Windows, iOS and Android)
- Support for Core ML, TFLite, ONNX, GGUF and more on-device formats
- Measure inference speed, load time, and peak RAM utilization
- Profile model layer execution across compute units (CPU, GPU, NPU)
- Up to 10 on-device model performance benchmarks overall
Pro
For indie developers
Pay As You Go
- Pre-paid benchmarking credits (1 credit = 1 benchmark)
- Credits never expire (use them anytime)
- Bulk credit packs at discounted rates
Team
For research and engineering teams
Custom
- Bespoke support with model conversion and optimization
- Evaluate model efficiency vs model accuracy trade-offs
- Priority access to devices for faster on-device performance results
- Weekly calls with RunLocal team and direct access on Slack
- Initial 1-month paid pilot
Trusted by
What Our "Team" Customers Say
"With their seamless model conversion service and outstanding technical support, RunLocal has reduced the time required for on-device model development from weeks to days. Their evaluation tooling makes it easy to assess model quality and performance at scale on different hardware and operating systems. RunLocal is a critical part of our model development and deployment process."
Backed by Investors
(and more)